[This version is provided by http://www.cyber-rights.org]
COM (97) 503 Communication from the Commission to the European Parliament, the Council, the Economic and Social Committee and the Committee of the Regions ensuring security and trust in electronic communication Towards a European Framework for Digital Signatures and Encryption - October 1997.
The need for secure electronic communication
Open networks such as the Internet are increasingly being used as a platform for communication in our society. Open and accessible, they allow rapid and efficient world-wide exchanges at low cost. This will lead to new forms of business configuration (e.g. "virtual" enterprises, work collaboration across the globe), of private communication (e.g. e-mail) and of organisation of public services (e.g. electronic tax declaration).
Open networks also have the capacity to offer substantial opportunities for global electronic commerce in goods and services which can be ordered, supplied and paid for electronically. Already today, software packages, information, music, and videos are being delivered over the Internet. It is now largely expected that electronic commerce will be one of the key drivers for the development of the global information society [ Communication of the Commission "A European Initiative in Electronic Commerce" (COM(97)157 final, 16.4.97)] .
Overall, the increasing use of open networks offers the possibility to create new businesses, new channels of distribution and new methods of reaching the customer. It also opens up opportunities to re-engineer business conduct itself.
However, the realisation of such developments are hampered by the noticed insecurities typical to open networks: messages can be intercepted and manipulated, the validity of documents can be denied, personal data can be illicitly collected. Fraud is already increasing in several forms. Therefore, today, important electronic documents are usually only exchanged in so-called "closed networks", that is, involving users between whom contractual relationships and mutual trust already exist. This model cannot be transferred to open networks because of the absence of such relationships between users. As a result, the attractiveness and advantage of electronic commerce and communication cannot be fully exploited.
In order to make good use of the commercial opportunities offered by electronic communication via open networks, a secure and trustworthy environment is therefore necessary. Cryptographic technologies are nowadays widely recognised as the essential tool for security and trust in electronic communication. Two important applications of cryptography are digital signatures and encryption. Digital signatures can help to prove the origin of data (authentication) and verify whether data has been altered (integrity). Encryption can help keeping data and communication confidential.
Several Member States announced their intentions to introduce specific regulation on cryptography and some have already done so. For example, Germany and Italy already moved ahead with digital signature laws. In other Member States internal discussions are taking place, and some tend to refrain, at least for the moment, from any specific regulation at all.
Divergent legal and technical approaches would constitute a serious obstacle to the Internal Market and would hinder the development of new economic activities linked to electronic commerce. An EU policy framework for ensuring security and trust in electronic communication and safeguarding the functioning of the Internal Market is therefore urgently needed. The European Union simply cannot afford a divided regulatory landscape in a field so vital for the economy and society.
As cryptographic services and products are more and more demanded, concerns are expressed that abuse of cryptography by criminals or terrorists would make it increasingly difficult to combat crime. Such concerns apply only to confidentiality services. Digital signatures do not pose any risk for law enforcement, since they do not prevent data from being read. Digital signatures could even bring significant law enforcement benefits as they allow for example messages to be attributed to a particular reader and/or sender. As, in addition, they need a specific regulatory framework to take into account their legal implications, the present Communication distinguishes between authentication and integrity services - digital signatures (part II) and confidentiality services - encryption (part III) [ This distinction is also stated clearly in the OECD Guidelines for Cryptography Policy, 27.3.97] .
In September 1996, the European Parliament invited the Commission to prepare legal EU provisions concerning information security and confidentiality, digital identification as well as the protection of privacy [ European Parliament Resolution A4-244/96, 19.9.96, OJ320, p.164, 28.10.96] . In November 1996 the Council of Ministers requested the Member States and the Commission to prepare consistent measures to ensure the integrity and authentication of electronically transmitted documents [ Council Resolution Nr. 96/C 376/01, 21.11.96 on new policy-priorities regarding the information society, OJ C376, 12.12.96] . In March 1997 the OECD adopted Guidelines for cryptography policy, setting out principles to guide countries in formulating their own policies related to the use of cryptography. These Guidelines - although non-binding - present the first attempt at international level to give policy orientations on several aspects of cryptography, including both encryption and digital signatures. The Bonn Ministerial Declaration of July 1997 also stressed the necessity of a legal and technical framework for digital signatures at European level as well as the importance of the availability of strong encryption technology for electronic commerce [ European Ministerial Conference, Bonn 6-8.7.97] .
In its April 1997 Communication on Electronic Commerce, the Commission
announced the intention to prepare a policy aiming at guaranteeing the
free movement of encryption technologies and products as well as to propose
a specific initiative on digital signatures. As announced the present Communication
aims at developing such a policy framework with a view to:
II. Authentication and Integrity: Digital Signatures
Transmitting data in electronic form has many advantages compared with traditional methods. Documents can be made available almost instantly and in any quantity and the recipient is able to work on them directly. Transmission is considerably cheaper and faster - documents can be sent around the globe in a matter of seconds, without delay. However, authentication and integrity services are needed for secure and trustworthy data transmission and communication over open networks.
The speed of technological progress implies that many of the potential application fields for authentication and integrity services are difficult to ascertain at this stage. New application areas (e.g. protection of intellectual property rights, stored data, network security or electronic cash) are developing continuously. In particular for electronic communication digital signatures are considered to play a significant role.
1. Digital signature: what it is and how it works
(i) Several different methods exist to sign documents electronically
varying from very simple methods (e.g. inserting a scanned image of a hand-written
signature in a word processing document) to very advanced methods (e.g.
using cryptography). Electronic signatures based on "public key cryptography"
are called digital signatures and widely considered as crucial for a variety
of applications [for a more detailed description see Annex
(iii) Contrary to cryptography used for confidentiality purposes, digital signatures are annexed to the data and leave the content e.g. of the signed electronic document or the electronic transaction intact. Of course, the data can in addition be encrypted as described and discussed in chapter III. The cryptographic technology is used to protect against the illicit use of signatures in an electronic environment. Technical means exist to signal when keys are being used for functionalities other then the one for which they have been generated (e.g. a key issued for authentication for confidentiality purposes).
(iv) With the help of the sender’s public key the recipient can find out whether the signed data has been altered and check that the public and private key of the sender are a complementary key-pair. Even the smallest change of the data would be discovered immediately. What appears to be a relatively complicated mathematical process is in practice carried out in a matter of seconds by the computer. The user therefore would not notice the underlying computing process.
(v) Verification of the authenticity and integrity of data does not necessarily prove the identity of the owner of the public key. How does for instance the recipient of a message know that the sender is really the one he claims to be? The public key may be attached to the message or be published in a directory, but what degree of confidence can the recipient have? Anyone can publish a public key under another name. The recipient may therefore wish to obtain more reliable information on the identity of the key owner. Such information can be given by the key owner himself, issuing the recipient with satisfactory proof. Another way is to have it confirmed by a third-party (e.g. a person or institution mutually trusted by both parties).
In the context of digital signatures these third-parties are most commonly so-called certification authorities.
2. Certification authorities (CAs)
The provision of public certification services is a completely new service sector. Although still in its infancy this sector is already raising a lot of interest. The sector is currently dominated by commercial undertakings based outside Europe, although some European companies have also emerged. A significant number of new entrants will appear on the market very rapidly. They seem to focus on their national market and do not, at least initially, target markets in other EU Member States. This hesitation is also linked to legal uncertainties.
CAs can perform a range of functions with regard to digital signatures. Sometimes, publications refer to them as Trusted Third Parties (TTPs). However, TTPs which in general may provide a wide range of services very often are perceived to stand for lawful access to encryption keys [see Annex III].
While it is not excluded that TTPs also act as a CA - as described in this paper - the functions of both institutions are considered to be different. In particular CAs are crucial for digital signatures to become a fully accepted tool within national legal systems, for instance, to ensure legal recognition and enforceability of a signature in electronic commerce. Therefore the role and the legal basis for CAs and TTPs need to be distinguished from a regulatory standpoint.
One central task of a CA is to authenticate the ownership and the characteristics of a public key so that they can be trusted. Once a CA is satisfied that the ownership and the characteristics of a public signature key are correct, a certificate is issued containing this key and other details. This certificate is itself digitally signed i.e. the CA signs the certificate with its private key to establish the correlation with the key owner. When the CA's public key is added, a simple automatic verification is possible. However, it is necessary for the recipient to trust the CA, in other words a CA must be mutually trusted by both parties.
As a result, several categories of certificates are technically conceivable, e.g. the CA’s public key can be signed by another CA leading to a certification hierarchy. It would also be possible to have the public key certified by several different CAs.
2.2. Possible contents of a certificate
A certificate can contain a whole range of informations, going beyond the mere key allocation and precisely determining its use. Some additional information will always be necessary, e.g. the algorithm to be used or the certificate expire date. Other information may be voluntary and will depend on the purpose for which the key is to be used and the level of confidence or trust required of it.
Examples of a certificate’s contents:
2.3. Key management
Key management implies an extensive task package, which can for instance include the generation and allocation of key-pairs, the identification of the owner, the creation of a public key directory and time stamping.
(i) Key creation and owner identification
The keys - which can also be generated by the user himself - must be effectively unique and tamper proof (which is practically given by the choice of an appropriate key length and generation procedure). Otherwise the digital signature cannot be allocated for legal relations in a reliable manner to data for which it has been generated and, via the key, to only one certain person or entity. This ensures that a key owner cannot refer to the fact that the digital signature was produced not with his key but with another one.
Keys may be allocated to private persons, legal persons (e.g. limited liability company) or to "entities without legal status" (e.g. department of an enterprise, working group). Keys can even be assigned to functional entities such as servers or PCs. Since the CA must guarantee the unique link between a key and its user, it has to identify the user in a reliable way and to hand out the key to the correct person.
(ii) Key directory
A directory of public keys may also be created providing information on the key owner, its validity period and other details, such as revocation. The key directory must always be kept up-to-date. Certificate revocation lists allow to determine whether a certificate has been revoked, suspended or reactivated. The effective operation of such a facility will depend on the speed and reliability of the cancellation procedure, which could be used in cases of invalidity of the certificate or loss and theft of the private key.
(iii) Time stamping
There are many situations in legal relations, where proof of the exact time of a certain action (transmission, creation or receipt of a document or the time at which a declaration of intent is made) is crucial. It is important to prove the exact time when a key was revoked to avoid liability for contracts signed with a compromised key. Therefore, digital time-stamping services able to reliably confirm the exact time of certain actions will be necessary. Time stamping services are also crucial for ‘Intellectual Property Right’ applications. These services could be provided by a CA, but of course also by another body.
2.4. Mutual recognition
In a fully international framework for electronic commerce certificates issued by foreign CAs must be mutually recognised in different countries. Thus the verification of any international certificate can be rapid and efficient. National structures could be complemented by a co-ordination mechanism at the European level. Such a concept is consistent with the Community’s established negotiation strategy on mutual recognition and could encourage the development of certification services in Europe. Agreements with third countries will be both easier to secure and economically more beneficial if done on the basis of a common Community-wide regime.
Mutual recognition provisions in national laws could in principle facilitate cross-border trust. They would at the same time reduce potential EU Internal Market obstacles and enhance crossborder circulation of goods and services. The direct application of the Treaty (Art. 30, 52, 59, mutual recognition provisions in national legislation) could already lead to a satisfactory functioning of the Internal Market. Other possibilities of ensuring cross-border recognition of certificates could be harmonised European certification services (including the procedures concerning the issuance of such a certificate) as well as common evaluation criteria and procedures.
Business partners sometimes do not have an interest in the precise identity of a particular person or entity, but only in the confirmation of previous contacts, in their affiliation to a defined group of persons, in their individual characteristics such as solvency and creditability or simply in unforged data.
Example: Credit card companies do not confirm the identity of the card-holder, but only whether this person has a certain line of credit.
Therefore in many cases people will have several key pairs corresponding to their different roles. Those persons not wishing or not obliged by law to communicate under their name can choose a pseudonym which safeguards their anonymity in transactions and communication (though the signatory is identified to the CA) whilst fully exploiting the integrity and authentication functions of digital signatures. This possibility is also required by the EU Data Protection Directive [ Directive 95/46/EC, 24.10.95 on the protection of individuals with regard to the processing of personal data and on the free movement of such data, OJ L 281/31, 23.11.95. See also Common Position 57/96, 12.9.96 with a view to the adoption of a European Parliament and Council Directive concerning the processing of personal data and the protection of privacy in the telecommunications sector, in particular in the integrated services digital network (ISDN) and in the public digital mobile networks, OJ C315, 24.10.96, which establish the specific rules for data protection and the right to privacy with regard to telecommunications networks. ] and supported by the OECD Cryptography Policy Guidelines. Without such a privacy safeguard, digital signatures could be abused as an efficient instrument for tracing individual on-line consumption patterns and communication or for intercepting, recording or misusing documents or messages.
There may be cases where the disclosure of pseudonyms may be necessary for reasons of public security and crime prevention. The EU Data Protection Directive lays down the conditions under which Member States may adopt measures restricting the right to remain anonymous.
Another privacy and data security concern results from the need that key pairs have to be unique and confidential in order to minimise the risks of "identity theft" and forgery. CAs must therefore be forbidden to store private keys. This again distinguishes CAs from TTPs which task is to keep information about private keys.
Since CAs must be able to identify the key owner and thus gather information about the individual, they are subject to the obligations concerning data processing, security and transfers to third countries laid down by the EU Data Protection Directive. For example, CAs can only collect and process personal data if the individual has given his consent or if they are authorised by law.
3. Legal Problems
While commercial products for digital signatures are already available in the market place, only a few companies in Europe have so far taken steps to offer services in this area. One of the main reasons is the weakness of demand resulting partly from the absence of legal recognition of digital signatures. Greater use of digital signatures requires adjustments and changes in many regulatory areas. In the current situation, the most important legal problems result from different national rules and regulations (or the lack of them), in particular the absence of common requirements for CAs, of technical and operational requirements to be met by certain categories of digital signature products, of liability rules and of legal recognition of digital signatures. The Commission will evaluate the possibility to provide for the harmonisation of the different national provisions to support international mutual recognition of digital signatures.
3.1.Elaborating Community requirements
At present there is no uniform legal framework specifying requirements for CAs in the European Union. This does not hinder CAs to be active on the market (there are already visible commercial activities in the US and also in the EU). But serious obstacles for cross-border trust would result from the lack of common rules.
Example: Certificates issued by a CA in one Member State would not be recognised by a CA in another Member State, especially if one Member State has foreseen a licensing system for CAs and the certificate has been issued by a foreign unlicensed CA.
Establishing common criteria for the activities of CAs on Community level would allow certificates issued by a CA in one Member State to be recognised in all other Member States (mutual recognition). Since these problems and the risk that divergent national rules, or the lack of such rules, will hold back the functioning of the Internal Market and the development of electronic commerce, there is a strong case for a common legal framework to be established. A Community framework would enhance trust in digital signatures, whilst promoting their legal recognition. Such a framework could for instance establish principles for the activities of CAs.
Example of fields where common requirements for CAs
could be specified:
It would also be essential to establish common technical requirements for digital signature products, if national provisions (e.g. for key generation or storage) will not be mutually recognised and hinder the functioning of the Internal Market. Community harmonisation measures should be limited to establishing the essential requirements and leaving technical details (e.g. through a mandate) to standardisation bodies.
Clear liability rules would contribute to the acceptance of CA services. However divergent levels of protection at national level could potentially act as a cross-border barrier to the provision of goods or services or to the use by public administrations of on-line services in a cross-border context. Liability questions may play a particular role in the relationship between users and CAs or between two CAs as well as with respect to licensing authorities (licensing CAs).
In all Member States, there are contractual rules connected to appropriate liability rules between the user and the CA. Liability depends very much on the concrete single cases. For instance, liability problems can be better managed if digital signatures are used within specific closed user groups.
Liability largely depends upon the concrete service offered by the CA as stipulated in the contract. A legal catalogue of requirements could form the basis for the contractual duties. It would also provide for both minimum and maximum liability of the CAs or guaranties, for example regarding the accuracy of the certificate or the correctness of the key directory. Certification practice statements, a detailed description of how certificate policies are implemented by a particular CA, could also play an important role as orientation for liability issues.
Normally there is no contractual relationship between a CA and third parties, like the recipient of a digitally signed message or another CA, who have confidence in the validity of certificates. Therefore Member States should examine whether there is a need for special liability rules.
Errors made by a licensing authority in the licensing process can be damaging to the user, the CA and third parties. Since the licensing authority has no contractual obligations and since the extra-contractual liability of public authorities is usually strictly limited, Member States should examine whether special rules for liability are necessary.
3.3. Legal recognition of digital signatures
The legal concepts behind signatures and the requirements on form and procedures, are different in each of the Member States jurisdictions. The differences, particularly in the field of civil and procedural law, have to be analysed. Member States should be encouraged to scrutinise the relevant national laws and regulations for provisions which do not allow to exploit the potential of digitally signed documents (form, evidence).
When signing a contract using a digital signature, one is confronted with different questions: does a declaration of intent have a legal value? Does the signature meet legal requirements? Is a digitally signed document recognised as evidence in court?
(i) Declarations of intent
Legal practices have emerged in Member States over the years in connection with declarations of intent. These cannot simply be translated into the context of electronic communication since the way to make a declaration of intent differs substantially from the traditional form in some respects.
Example: The delivery of a document in paper form requires more time than in the electronic form. One has to put the document into an envelope, apply a postage stamp and post it. In so doing, one still has time to reconsider one’s decision. An electronic document on the other hand is delivered by simply pressing a key or button.
In particular in order to guarantee an appropriate protection against hasty decisions, Member States should examine whether specific requirements are needed regarding the binding character of declarations of intent.
In addition, technical solutions must be found to make sure that users sign a document in the version which is actually visible on their screen.
Example: Technically, substantial differences may exist between the document visible on the screen and the document which is actually signed or printed, e.g. if the programme works with associated files.
(ii) Non-repudiation of digital signatures
Even when a key pair has been assigned in total trust to a certain person, this does not prove that this person has actually signed a given document. While the normal situation is that the key owner signs the document, a digital signature can in fact only be associated with certainty to a given private key. This presumption will only hold if it is certain that only the owner of the secret private key has full and unique control over his private key. Key escrow of private keys would endanger this presumption.
Example: Unlike conventional signatures, where the signatory signs with his own hand, digital signatures also allow a third - authorised or unauthorised - person to sign the document if this person is in possession of the private key, so-called "undisclosed" delegation.
Assignment is however possible if it can be legally presumed that the key owner signed himself. In that case the owner might wish to be legally liable only to a certain extent (e.g. within a limit, as with a credit-card). Member States should therefore consider appropriate legal rules.
(iii) Legal treatment of references
In order to carry business transactions faster or for cost reasons, one can refer to documents which are not part of the electronically transmitted data itself, but which are stored in another place, e.g. reference to standard-form contract conditions, technical descriptions or plans.
Problems could however result from the fact that the technical possibility of referring to other documents does not meet the legal requirements that have emerged from traditional legal relations.
Example: In a sales contract, a computer company refers to the terms of delivery indicated on the company’s Internet-homepage. Under which conditions do the terms of delivery become part of the contract? Do they have to be digitally signed as well?
Special rules in Member States’ civil laws will therefore be necessary for the legal treatment of references in electronic legal relations. The most important point is that references do not have other legal effects than those they would have if they were contained in the document in question.
(iv) Legal effects
Ensuring equivalent legal effects for conventional hand-written and digital signatures is not easy to realise considering their different characteristics and their different ways of being materialised.
Member States could also implement specific rules on an electronic form in their civil laws. Thus Member States would not have to change all their regulations on written form but would be able to introduce digital signatures only where they think it would make sense.
Legal domains in which no specific legal form is prescribed, but where, for example, the use of the written form is based on voluntary business practice, would greatly benefit in terms of security - thanks to the gain of confidence - from the legal recognition of digital signatures.
4. Regulatory considerations
(i) While digital signatures are currently a recognised answer to authentication and integrity questions, the market may come up with other solutions. Therefore regulation has to create on one side a clear framework to build trust in digital signatures, but on the other side also has to be flexible enough to react to new technological developments.
(ii) Regulation should not restrict, neither de jure nor de facto, the contractual freedom of parties. Therefore any regulation should be tailored to correspond to the different possible uses of digital signatures (see II.1.). Private use of digital signatures or use within closed-user groups, for instance, might escape specific regulation entirely. Well-identified cases could become subject to regulation, for example in official communication. In any case, it must be ensured that both regulated and unregulated digital signature schemes can co-exist and are interoperable.
(iii) Some Member States are in the process of introducing voluntary schemes, and others consider mandatory licensing schemes, to build trust in CAs and to encourage legal recognition of digital signatures. However, licensing is only one of the possible trust-enhancing methods Member States may apply to promote the use of legally valid digital signatures. Non-licensed, but highly regarded private or public organisations may as well be considered as a trusted CA.
(iv) In the context of licensing, it is important to distinguish clearly between on the one hand, the procedures and conditions governing the establishment of a CA, and, on the other hand, the conditions imposed on the different services provided by a CA. The Treaty Articles 52 and 59 apply to each of these situations. Different national regulatory approaches and the lack of mutual recognition of each other’s regulatory requirements may easily lead, due to the inherent cross-border nature of digital signatures, to a fragmentation of the Internal Market for electronic commerce and on-line services throughout the Union.
(v) Restrictive practices with regard to the establishment of CAs, the services they provide, the cryptographic tools they use, etc. will be detrimental to the free circulation of goods and services within the Internal Market. They should not undermine the freedom of establishment, for example by discriminating without justification on the basis of nationality or by restricting without justification the number of those providing CA services. The scope and the timeframe of Community action would be determined by the need for harmonisation. Since mandatory licensing of CAs is not the only way to ensure compliance of CA’s activities with public intentions of how to promote trust in digital signatures, an EU regulatory framework would have to provide for the co-existence of both licensed and unlicensed CAs. Such a framework should be put in place at the latest by the year 2000.
III. Confidential electronic communication: Encryption
1. The economic and societal importance of encryption
(i) An encryption algorithm transforms a plaintext into an unreadable ciphered text (encryption) and vice versa (decryption) using a special key. The economics behind encryption is to transform the problem of keeping thousands of messages secret into the problem of keeping a single key secret. A useful distinction can be made between symmetric and asymmetric encryption algorithms [see Annex II for more detailed explanation].
Symmetric algorithms use the same key for encryption and decryption. This means that communicating parties have to agree on a secret key in advance. The disadvantage is that they have to find a secure way to exchange this key. This is particularly cumbersome in an open environment with many participants that may not know each other beforehand. This disadvantage is avoided in asymmetric encryption methods that use different keys for encryption and decryption.
At present, encryption provides the most important tool to keep electronic communication and electronically stored documents confidential. Although new technologies will emerge sooner or later, it can be expected that encryption will remain the cornerstone for most confidentiality services on open networks for the foreseeable future.
Encryption has a long tradition in the defence area. However encryption technologies are increasingly integrated into commercial systems and applications.
Computer stores sell cryptographic products and more and more people simply down-load encryption software from the Internet which can be easily installed on a normal PC. The integration of complete cipher machines on smart cards is a reality. PCs could be delivered with standardised smart card readers and fast crypto-chips. Various universities in the world teach cryptology and hundreds of companies in Europe and even more world-wide develop, produce and sell products and systems to be used for encryption.
A survey has identified not less than 1,400 encryption computer products world-wide [ Survey conducted by Trusted Information Systems] . More than 400 companies from the US and about 440 companies outside the US, many of them in Europe, now offer encryption products [ see also Computer Systems Policy Project CSPP: "Perspectives on security in the information age ", January 1996. CSPP is an affiliation of chief executive officers of leading American computer companies] . Involved in this process are incumbents like computer, software and telecommunication companies as well as high-tech start-ups. Most of the young companies are growing fast: numerous examples exist where the annual growth rates of turnover or employment are 100% and even more.
(iii) Electronic commerce and many other applications of the information society will only expand and unfold their economic and social benefits if confidentiality can be assured in a user-friendly and cost-efficient way.
Furthermore, the application of cryptographic products and services will have an enabling effect in all sectors of economic and social activity. Without this widescale deployment, the ability to create new, more competitive forms of business and new forms of social interaction will be substantially inhibited.
(iv) International treaties, constitutions and laws guarantee the fundamental right to privacy including secrecy of communications [ Art. 12 Universal Declaration of Human Rights, Art. 17 International Covenant on Civil and Political Rights, Art. 8 European Convention on Human Rights, Art. F(2) Treaty on EU, EU Data Protection Directive] . Consequently, in the current shift from off-line to on-line information flows, the public needs to have access to technical tools allowing effective protection of the confidentiality of data and communication against arbitrary intrusions. Encryption of data is very often the only effective and cost-efficient way of meeting these requirements. Therefore, the debate about the prohibition or limitation of the use of encryption directly affects the right to privacy, its effective exercise and the harmonisation of data protection laws in the Internal Market.
2. Regulation of encryption: Potential impact on the Internal Market
2.1. Export control measures
Concerns over foreign threats to national security have been the primary motive for export controls. Whilst countries want to protect their own military and diplomatic communication through encryption, the objective of export control is precisely to deny similar benefits of cryptography to foreign opponents, in particular if they do not have equivalent technical means. Therefore, export controls are in general designed to prevent international proliferation of certain encryption technologies.
Under the Wassenaar arrangement on export controls for conventional arms and dual-use goods and technologies (19.12.1995) [ see http://www2.nttca.com:8010/infomofa/press/c_s/wassenaar.html; http://ideath.parrhesia.com/ wassenaar/wassenaar.html] , replacing the COCOM [ Co-ordinating Committee for Multilateral Export Controls was an international organisation for the control of the export of strategic products and technologies to proscribed destinations. Members were to a large extent NATO countries but also others like Japan and Australia.] list, a group of 28 countries apply export controls to encryption products.
Within the European Union, the Dual-Use Regulation of December 1994 establishes a common framework for exports of dual-use goods [ Council Regulation (EC) 3381/94, 19.12.94 setting up a Community regime for the control of exports of dual-use goods, OJ L 367/1, 31.12.94. Council Decision 94/942/CFSP, 19.12.94 establishes the lists of dual-use goods covered by the Regulation, OJ L 367/8, 31.12.94.] . Certain encryption products may only be exported on the basis of an authorisation. In order to establish an Internal Market for dual-use goods, such export authorisations are valid throughout the Community.
Moreover, according to Article 19 of this Dual-Use Regulation, Member States exercise a licence procedure for a transitional period also for intra-Community trade for certain particularly sensitive products. For the time being this also includes encryption products. This means the Regulation obliges Member States to impose not only export controls (i.e. controls on goods leaving Community territory) on dual-use goods, but also intra-Community controls on cryptography products shipped from one Member State to another.
The Dual Use Regulation however does not fully specify the scope, content and implementation practices of national controls. Consequently, a large variety of domestic licensing schemes and practices exists. These divergences can lead to distortion of competition.
2.2. Domestic control measures
Law enforcement authorities and national security agencies are concerned that wide-spread use of encrypted communication will diminish their capacity to fight against crime or prevent criminal and terrorist activities. For this reason, in several Member States consideration is being given to how their encryption policy could develop in the future. This has led to national and international discussions about the need, technical possibilities, effectiveness, proportionality and privacy implications of such a regulation.
(i) Existing regulation within the European Union and the OECD
Whilst export control measures are internationally widely applied, up to now, domestic control of encryption is quite exceptional. In fact, currently only one Member State of the European Union (France), applies a comprehensive cryptographic regulation [ Loi N° 90-1170 of 29.12.90, JORF 30.12. 90; Decret N° 92-1358, 28.12.92, JORF 30.12.92 ;. Delivery, exportation and use of cryptography are subjected to previous declaration if the cryptography can have no other object than authenticating communications or assuring the integrity of transmitted messages, and previous authorisation by the Prime Minister in all other cases. This law is currently being modified according to loi N° 96-659, 26.7.96 de réglementation des télécommunications art 17] . Although there have been discussions in other Member States, only the United Kingdom has so far launched a Public Consultation on the regulation of TTPs for the provision of encryption services (but not for use of encryption) [ Licensing of TTPs for the provision of encryption services - DTI Public Consultation Paper on detailed proposals for legislation, 3.1997] .
The international picture is quite similar. Looking at the OECD countries, besides export controls there are basically no domestic regulations implemented. In the US - where up to now no domestic regulation is in place - there is an intensive debate on several legislative initiatives. In taking up the developing debate on this topic in some OECD Member countries and trying to avoid obstacles to international trade and commerce resulting from divergent national policies, the OECD has adopted Guidelines for a cryptography policy.
(ii) Regulation of use of encryption
Regulation of use would mean to rule the use of encryption without an authorisation as illegal. Alternatively or additionally, supply and import of encryption products and services could be brought under an authorisation scheme. Authorisations would either be denied or granted under certain conditions, for instance to use only weak encryption or to sell only approved software. These conditions are scaleable to satisfy any perceived needs of law enforcement and national security agencies.
Such regulations could limit the use of encryption. In addition, divergence between regulatory schemes might result in obstacles to the functioning of the Internal Market, in particular for the free circulation.
If an encryption software company which can freely develop its products in its home country, must comply with specific technical or legal requirements in other Member States, this company has to produce at least two, if not more, different versions of its encryption software. The same situation occurs if enterprises want to offer cross-border encryption services.
Today, nobody can be totally prevented from encrypting data (criminals or terrorists also can use encryption for their activities [ Most of the (few) criminal cases involving encryption that are quoted as examples for the need of regulation concern "professional" use of encryption. It seems unlikely that in such cases the use of encryption could be effectively controlled by regulation; see also "Encryption and Evolving technologies as tools of organised crime and terrorism" by D.E. Denning and W.E. Baugh, Jr. ] ): Firstly, access to encryption software is relatively easy, for instance by simply downloading it from the Internet. Secondly, it is difficult to prove that a specific person has sent an unauthorised encrypted message. Electronic communication on open networks is not like an end-to-end telephone conversation where people can be identified for instance by their voice. Thirdly, encryption is also possible using steganographic methods [ see Annex II] . These methods allow one to hide a message in other data (e.g. images) in such a way that even the existence of a secret message and thus the use of encryption cannot be detected.
As a result, restricting the use of encryption could well prevent law-abiding companies and citizens from protecting themselves against criminal attacks. It would not however prevent totally criminals from using these technologies.
2.3. Lawful access to encryption keys
The underlying principle of this approach is to require that products and services incorporating encryption allow access to the respective keys. This would permit government agencies to decrypt a ciphered text otherwise difficult or impossible to crack. Different technical and institutional ways to provide key access are being discussed. The two most known concepts are key escrow and key recovery. Broadly speaking, these concepts imply that copies (escrow concept) or information (recovery concept) about relevant keys are given either directly to government agencies or to TTPs [see Annex III].
(i) Key access schemes are considered by law enforcement agencies as
a possible solution to cope with issues like encrypted messages. However
these schemes and associated TTPs raise a number of critical questions
that would need to be carefully addressed before introducing them. The
ongoing discussion of different legislative initiatives in the US is an
illustrative example of the implied controversy. The most critical points
are vulnerability, privacy, costs and effectiveness:
Users may therefore not see many advantages in using TTPs for confidential communication, and probably not even for stored information. Regulators would thus need to offer incentives to convince users to use licensed TTPs for confidentiality purposes, for instance through a "public security label" or even by introducing a "mandatory scheme". Such a mandatory scheme would make any publicly available offer of encryption services subject to a licence that inter alia would demand key escrow/recovery.
The acceptance of such a system remains to be seen, but given its implied overheads, can not be regarded as an incentive for electronic commerce. In any case, restrictions imposed by national licensing schemes, particularly those of a mandatory nature, could lead to Internal Market obstacles and reduce the competitiveness of the European Industry.
Privacy considerations suggest not to limit the use of cryptography as a means to ensure data security and confidentiality. The fundamental right of privacy has to be ensured, but may be restricted for other legitimate reasons such as safeguarding national security or combating crime, if these restrictions are appropriate, effective, necessary and proportionate in order to achieve these other objectives. The EU Data Protection Directive harmonises the conditions under which access to personal data, their processing and transfer to third countries is lawful.
As regards data security the Directive requires Member States to provide that a data controller must implement appropriate technical and organisational measures to protect personal data against accidental or unlawful destruction or accidental loss, alteration, unauthorised disclosure or access, in particular where the processing involves the transmission of data over a network, and against all other unlawful forms of processing.
Cryptography is one important technical means by which data integrity and their confidentiality can be ensured. To ensure also the secure flow of personal data throughout the Internal Market, such technical means must be able to "travel" with the personal information they are securing. Any regulation hindering the use of encryption products and services throughout the Internal Market thus hinders the secure and free flow of personal information and the provision of related goods and services.
Proposals for regulation of encryption have generated considerable controversy. Industry expresses major concerns about encryption regulation, including key escrow and key recovery schemes [ see e.g. Industrial Declaration of the Bonn conference, July 97] . Although there is a lack of experience, as electronic communication and commerce have just begun to penetrate economy and society, a preliminary assessment can be made in order to build a common European understanding of the subject, in particular as Member States may have different views on security issues implied. Such an understanding could be founded on the following points:
(i) Problems caused by encryption to crime investigation and the finding of evidence are currently limited, but they may increase in the future. As with any new technology, there will be abuse of encryption and criminal investigations will be hindered because data was encrypted. However, widespread availability of encryption can also prevent crime. Already today, the damage caused by electronic crime is estimated in the order of billions of ECUs (industrial espionage, credit card fraud, toll fraud on cellular telephones, piracy on pay TV encryption). Therefore, there are considerable economic and legal benefits associated with encryption.
(ii) Criminals cannot be entirely prevented from having access to strong encryption and from bypassing escrowed encryption. Benefits of regulation for crime fighting are therefore not easy to assess and often expressed in a fairly general language. However control measures could make use of encryption for criminal activities more difficult and cumbersome.
(iii) In the information society, citizens and companies will increasingly carry out more aspects of their lives and business on-line. Through teleconferencing, tele-shopping, teleworking, electronic payment, e-mail, etc. a huge amount of information will be available electronically, in a way never experienced before. Therefore, if citizens and companies have to fear that their communication and transactions are monitored with the help of key access or similar schemes unduly enlarging the general surveillance possibility of government agencies, they may prefer remaining in the anonymous off-line world and electronic commerce will just not happen [see Eurobarometer opinion survey 46.1 on privacy in the information society, January 1997] .
(iv) Key escrow or key recovery raise a number of practical and complex questions that policy makers would need to solve, in particular issues of privacy, vulnerability, effectiveness and costs. If at all required, regulation should be limited to what is absolutely necessary. Regulation would also need to distinguish between a multitude of possible key types (storage keys, session keys, authentication keys, etc.) as there are important differences in their functionality.
(v) In the context of electronic commerce using open and global networks, the international availability, interoperability and choice of various encryption products and services is necessary. Any regulation hindering the use of encryption products and services throughout the Internal Market hinders the secure and free flow of personal information and the provision of related goods and services, and its justification needs to be examined in light of the Treaty and the EU Data Protection Directive
(vi) The ultimate objective for government agencies is to see plaintext and not necessarily to have access to keys. Furthermore traffic analysis (e.g. who communicates with whom?) is also important and would benefit from increased electronic communications. Information, even encrypted for communication, can often be found unencrypted at the source, just as with traditional forms of communication, for instance with banks, shops, travel agencies involved in communication with a suspect, or can be tapped unencrypted at certain points in a communication link. Therefore existing regulation on traditional forms of lawful access to data and communication could be explored with a view to effectively applying it to access to encrypted data and communication, e.g. regulation could require access provision to encrypted information upon legally authorised request.
(vii) A fundamental problem lies in international relations, i.e. how to ensure global communication in case key escrow/recovery regulation is introduced in some countries. Countries would probably insist that only national TTPs could hold keys of their citizens. For instance, in case of a session key recovery scheme that is linked to an e-mail communication, only the country of the sender could decrypt the message unless there is a special arrangement between the two countries.
(viii) Irrespective of the compatibility of restrictions with the Treaty provisions on the free circulation of goods and services, specific national controls on the use of encryption could also have a secondary effect on the free circulation of persons, similar to those already identified by the Veil Panel [ Report of the High Level Panel on the free movement of persons, chaired by Mrs. Simone Veil, presented to the Commission, 18.3.97] .
IV. Policy actions at Community level
Electronic communication via open networks is at the core of the information society. Fast and secure exchange of data offers many advantages for electronic commerce which can contribute decisively to improvements in competitiveness and job creation. The European Union has an early opportunity to create the conditions for a trend-setting infrastructure and for growth in European industry.
The Commission will seek to build trust in electronic communication via open networks to ensure the functioning of the Internal Market, to stimulate electronic commerce and to strengthen the European Industry.
1. Community framework for digital signatures
1.1. The need for European Union action
Detailed regulations for digital signatures are already under preparation in some Member States. France has already adopted a new Telecommunications Act, Germany a law on digital signatures [ Gesetz zur digitalen Signatur (SigG), 1.8.97] , Italy a law on the use of electronic documents and contracts [ Schema di Regolamento "Atti, documenti e contratti in forma elettronica", approved by the Italian Council of ministers, 5.8.97] . The UK Government has launched a Public Consultation on the regulation of TTPs. The Dutch Government has created an inter-departmental task force [ Staatscourant nr. 54, 18.3.97] . Denmark and Belgium [ see http://www.agoraproject.org/] are also preparing draft legislation on digital signatures. The Swedish government organised a public hearing in June 1997.
Whilst the development of a clear framework is welcomed, the very divergent legal and technical approaches which have already appeared and the absence of any legal environment in other Member States - also possibly justified - might constitute a serious barrier to doing business and communicating throughout the European Union. This will undermine the free circulation of digital signature related products and services within the Internal Market as well as the development of new economic activities linked to electronic commerce. In order to stimulate electronic commerce and the competitiveness of the European industry as well as to abolish the free circulation obstacles and to facilitate the use of digital signatures across national borders, a common framework at Community level is urgently needed and should be put in place at the latest by the year 2000.
1.2. Scope of a Community framework
The goal of any Community initiative must be to encourage Member States to rapidly implement appropriate measures to build trust in digital signatures. The Commission therefore considers proposing - in the context of the Amsterdam Treaty - first pillar legislation on the basis of this Communication. The following steps would be necessary from the Commission’s point of view:
(i) Common legal requirements for CAs
Common European certification requirements are crucial. By establishing defined common criteria for the activities of CAs, the Community could put in place a framework allowing that certificates issued by a CA in one Member State are recognised in all other Member States. A Community framework would have to refer particularly to the setting of common requirements for the establishment and operation of CAs allowing for the co-existence of licensed and non-licensed CAs. Common classes of certificates may also be needed so that the levels of assurance and trust for certificates are the same in all Member States. Detailed implementation and the means of applying such rules (licensing regime, self-certification) would be a matter for Member States to decide.
To support international mutual recognition of digital signatures the Commission will furthermore identify the need for common technical and operational requirements as well as common evaluation criteria and procedures, including standards, concerning digital signature products.
(ii) Legal recognition
In order to achieve as wide as possible acceptance of digital signatures, national legal systems may need to be adapted to ensure that they offer the same recognition and treatment to digital signatures as to conventional signatures.
The Commission will complete its currently ongoing assessment of the need to provide for the legal recognition of digital signatures at Community level. The different national provisions inhibiting the full exploitation of digitally signed electronic documents (form requirements, evidence rules), on the basis of which further proposals for action will be made will also be taken into account. Legal form requirements and the validity of signatures as evidence in legal proceedings should rapidly be submitted to examination by justice ministers.
(iii) International co-operation
Electronic communication is not limited to the European Union. Therefore - where appropriate - a framework must be developed at an international level once a Community position has been established. This requires participation of Europe (both on Community and on Member States level) in international initiatives and fora.
Many international initiatives have been initiated at different levels. Bilateral (EU/US, EU/Japan) and multilateral (e.g. UNCITRAL [ United Nations Commission on International Trade Law] ) discussions have started. UNCITRAL has completed the work on a Model Law on Electronic Commerce and has recently initiated subsequent work aiming at the preparation of uniform rules on digital signatures and the related (cross-border) services (CAs). Work in the OECD based on the Guidelines for cryptography policy is continuing. Other international organisations, such as the WTO, may become involved with regard to avoiding trade obstacles and other aspects related to their specific area of competence and expertise.
In the United States [ An update on the status of US legislation can be found on http://www.mbc.com/ds_sum.html] almost all States have either started working on or have already legislation on digital signatures. Agencies, such as the Food and Drugs Administration, are promulgating regulations specific to their area of responsibility [ http://www.fda.gov/cder/esig/part11.htm] . At the federal level, Congress is considering several legislative initiatives. In Japan, some technical and regulatory activities in the area of authentication and electronic transactions have been launched earlier this year.
At the business level the American Bar Association produced the "Digital Signature Guidelines" and the Internet Law and Policy Forum (ILPF) is working on the role of CAs in consumer transactions.
In view of these world-wide activities the Commission recommends that the Community continues and initiates the dialogues on international level. The goal must be to remove existing obstacles in order to create an internationally compatible framework for electronic commerce, in particular to establish common technical standards and mutual recognition of certificates.
2. Policy orientations in the area of encryption
(i) The EC Treaty and the Treaty on the European Union fully respect
the competence of Member States with regard to the areas of national security
and law enforcement. If national restrictions are put into place they have
to be compatible with Community law. Therefore the Commission will examine
whether national restrictions are totally or partially justified, notably
with respect to the free circulation provisions of the Treaty, the case
law of the Court of Justice and the requirements imposed by the Data Protection
(ii) The Dual-Use Regulation should be adapted in view of the requirements
for the cryptographic products market. Article 19 imposing national controls
also contains a provision to re-examine the need for these controls within
three years from the date of entry of the Regulation (by the end of 1997).
Therefore, when the Dual-Use Regulation is reviewed it could be improved
(iv) Given the global dimension of electronic communication and commerce, international agreements may be necessary between the Community and other countries, once a harmonised system has been put in place. The goal must be to remove existing obstacles in order to create an internationally compatible framework for electronic commerce, in particular to establish common technical standards and mutual recognition of certificates.
(v) The Council is also invited to initiate a debate on encryption issues.
3. Accompanying measures
Interoperability between different encryption and digital signature applications and systems is absolutely necessary to ensure that they can be applied in and outside Europe. Services are mostly achieved by agreed standards including test criteria and procedures covering protocols, data formats and program interfaces.
By using agreed protocols and data formats it is not necessary to develop gateway services or conversion programs changing one format to another. Interoperability in a broader sense also means that application solutions can be moved from one type of software and hardware environment to another (portability) and that users can move from one place to another and still access the same trusted services (mobility).
Examples for work on standards:
The Commission encourages industry and international standards organisations to develop technical and infrastructure standards for digital signatures and encryption to ensure secure and trustworthy use of networks and respect privacy and data protection requirements [ see Bonn Ministerial declaration, footnote 5] . The Commission will consider specific mandates on standardisation and will propose, in close co-operation with the Member States, industry as well as the user community (business, consumers, citizens) measures which will support the work in this field.
(ii) Support programme
The Commission is ready to support the development of cryptographic services, in particular it is considering proposing a Council and Parliament Decision for an INFOSEC II programme building on the INFOSEC programme carried out from 1992 until 1994. The programme could aim at developing overall strategies for the security of electronic communication, in particular with a view to provide users and producers of electronic communication with appropriate protection systems.
(iii) Research projects
The Commission will continue the current projects in the field of digital signatures and encryption within the 4th framework programme for Community activities in the field of research and technological development (1994 - 1998) [see Annex IV for a list of ongoing projects] and will launch new projects within the 5th framework programme (1998 - 2002). Notably the proposal for the 5th framework programme foresees a key action on electronic commerce. Special importance will be attached to techniques aiming at interoperability and enhancing privacy, to stimulating best practice and encouraging its widescale deployment.
(iv) The use of digital signatures and encryption by public authorities
In the near future, government administrations will use digital signatures and encryption for internal purposes or in their relations with business and citizens. Such use may require adaptations to national as well as Community laws, regulations and administrative procedures. The first Community Regulation [ Council Regulation (EC) N° 1290/97, 27.6.97 amending Regulation (EEC) N° 1408/71 on the application of social security schemes to employed persons, to self-employed persons and to members of their families moving within the Community and Regulation (EEC) N° 574/72 laying down the procedure for implementing Regulation (EEC) N° 1408/71 OJ L 176, 4.7.97, P. 1 insertion of a new paragraph in Article 85 ensuring that documents exchanged by electronic means are given the same status as paper documents] has been modified in order to allow the use of digitally signed electronic documents. The impact of national measures has to be monitored in order to identify problem areas which may require a Community intervention. Also the Union’s institutions will use digital signatures [ SINCOM, the budget management application of the Commission, introduces smart cards for digital signatures purposes] and encryption.
(v) European Internet-Forum
The Commission will create by the end of 1997 an electronically based European Internet-Forum as a means to exchange information on the regulatory and user aspects of digital signatures and encryption.
(vi) International hearing
The Commission intends to organise beginning of 1998 a hearing about
the topic "digital signature and encryption". The aim is to consult governments,
industry and consumers on which measures they feel the Community should
take into consideration in order to
4.Q./1997: European Internet-Forum
4.Q./1997: Commission proposal to amend the Dual-Use Regulation
1.Q./1998: International hearing
1.Q./1998: Assessment of the comments on the Communication, the results of the Internet-Forum and the international hearing
2.Q./1998: Proposal for further action (e.g. Directive on digital signatures)
2.Q./1998: Proposal for an Infosec II programme
1998-2002: Projects within the 5th framework programme by 2000: Common framework for cryptography put in place throughout the Union